In the life sciences, fluorescence is typically used as a non-invasive method of identifying and analyzing biological materials. Specific targets in the biological material such as for example, proteins, nucleic acids, lipids, cells and cell components, stem cells or small molecules can be labeled with an extrinsic or exogenous fluorophore, and thus subsequently imaged. Biological materials also naturally fluoresce, which is known as intrinsic fluorescence or “autofluorescence” because it occurs in the absence of exogenously administered fluorophores. Autofluorescence is believed to originate from various endogenous fluorophores in biological materials, including for example nicotinamide adenine dinucleodde (NADH), elastin, collagen, flavins, amino acids and porphyrins.
Autofluorescence and fluorescence emission can be generated and recorded as images when light with the appropriate excitation wavelengths illuminates the biological material. However, autofluorescence, which is the result of a combination of fluorophores and is characterized by broad emission spectra extending over several hundred nanometers, can interfere with the ability to detect the emission of a specific fluorophore, when the emission spectra of the fluorophore and the autofluorescence overlap. In such instances, in addition to reducing signal detection sensitivity by masking the fluorescence of the fluorophore of interest, autofluorescence may also decrease the specificity of detection by providing false positive results.
One approach to addressing this problem is to utilize means to reduce or minimize the detected emission signal that is contributed by autofluorescence of the biological material. The prior art describes methods to reduce autofluorescence by employing various pre-treatments of the biological material prior to image acquisition. However, such techniques may also degrade the quality of the biological material itself, and are typically not suitable for in vivo applications. Alternatively, if the autofluorescence emission itself cannot be mitigated, it is possible to minimize the contribution of signal from autofluorescence to image data by means of digital manipulation of any acquired fluorescence images. For example, in images containing the combined signal from both the fluorophore of interest and autofluorescence, some of these methods rely on acquiring estimates of the “pure” autofluorescence signal and using such estimates to remove autofluorescence by a weighted subtraction. Other methods use statistical correlation techniques to correct for the additive autofluorescence signal. These image data manipulation techniques are described in prior art references and are generally limited by poor accuracy, by the need for small (i.e., low resolution) data sets, or by the need for significant post-processing. It is consequently desirable to establish a high resolution image processing technique to quickly and accurately distinguish the fluorescence emitted by a fluorophore of interest in a biological material from the autofluorescence emission in that same biological material.